Re: Sequential scan on FK join

Поиск
Список
Период
Сортировка
От Alvaro Herrera
Тема Re: Sequential scan on FK join
Дата
Msg-id 20051022132512.GE18064@surnet.cl
обсуждение исходный текст
Ответ на Re: Sequential scan on FK join  (Martin Nickel <martin@portant.com>)
Список pgsql-performance
Martin Nickel wrote:
> On Tue, 18 Oct 2005 08:52:15 +0100, Richard Huxton wrote:
> >   3. Actually - are you happy that your general configuration is OK?
> We're running dual Opteron 244s with 4G of memory.  The platform is
> Suse 9.3, 64 bit.  The database is on a 3ware 9500S-8 sata raid controller
> configured raid 10 with 4 drives plus a hot swap.  Drives are
> 7400 rpm (don't remember model or size).
>
> I'm running Postgres 8.0.3.  Here are some of the relevant conf file
> parameters:
> shared_buffers = 50000
> sort_mem = 8192
> work_mem = 256000

Interesting that you set both sort_mem and work_mem.  Do you realize
that the former is an obsolete name, and currently a synonym for the
latter?  Maybe the problem is that you are using too much memory for
sorts, forcing swap usage, etc.

--
Alvaro Herrera                 http://www.amazon.com/gp/registry/DXLWNGRJD34J
"La persona que no quería pecar / estaba obligada a sentarse
 en duras y empinadas sillas    / desprovistas, por cierto
 de blandos atenuantes"                          (Patricio Vogel)

В списке pgsql-performance по дате отправления:

Предыдущее
От: "Karl O. Pinc"
Дата:
Сообщение: Using LIMIT 1 in plpgsql PERFORM statements
Следующее
От: Kishore B
Дата:
Сообщение: Need help in setting optimal configuration for a huge database.